Reformat covariates to use in model

athowes / malaria-decline-us

Factors influencing the eradication of malaria in the United States 1920-1950

0 stars 0 forks source link

Reformat covariates to use in model #1

Closed athowes closed 2 years ago

athowes commented 3 years ago

Move from wide format covariate_year to long format c(year, covariate)
Might involve fitting models to intraplate them into years not available

athowes commented 3 years ago

https://arxiv.org/abs/1912.10981
https://becarioprecario.bitbucket.io/inla-gitbook/ch-missing.html#sec:misscovar

athowes commented 3 years ago

Use missForest algorithm for imputation
Plot missing data with e.g. vis_miss()

athowes commented 3 years ago

Justin and Abby think that they would be able to get more coverage of some of the covariates. Still might be the case that many of them are imputed though.